You are on page 1of 3

Journal of Experimental Social Psychology 46 (2010) 226–228

Contents lists available at ScienceDirect

Journal of Experimental Social Psychology


journal homepage: www.elsevier.com/locate/jesp

FlashReport

When the death makes you smoke: A terror management perspective


on the effectiveness of cigarette on-pack warnings
Jochim Hansen a,b,*, Susanne Winzeler b, Sascha Topolinski c
a
New York University, Department of Psychology, 6 Washington Place, New York, NY 10003, United States
b
University of Basel, Department of Psychology, Missionsstrasse 60/62, CH-4055 Basel, Switzerland
c
University of Würzburg, Lehrstuhl für Psychologie II, Röntgenring 10, D-97070 Würzburg, Germany

a r t i c l e i n f o a b s t r a c t

Article history: One of the principal vehicles for informing tobacco consumers about the risks of smoking is the warning
Received 16 July 2009 message on each cigarette package. Based on terror management theory, the present study investigates
Revised 11 September 2009 the impact of mortality-salient warnings on cigarette packages compared to warnings with no mortality
Available online 18 September 2009
threat. Results suggest that to the degree that smoking is a source of self-esteem, later attitudes towards
smoking become more positive if the warning message is mortality-salient. On the contrary, if the warn-
Keywords: ing is terrifying but not mortality-salient and relates to the source of self-esteem, smoking attitudes
Terror management health model
become more negative with higher smoking-based self-esteem. Thus, mortality-salient warnings may
Terror management theory
Smoking
increase the tendency to favor smoking under certain circumstances. This fatal ironic effect highlights
Warning message the importance of a risk communication that matches the self-esteem contingencies of the recipients,
Mortality salience and it has urgent implications for health care policy.
Self-esteem Ó 2009 Elsevier Inc. All rights reserved.

Smoking is responsible for millions of deaths all over the world; pose that, on the one hand, terrifying but death-unrelated warn-
and the figure is estimated to continue to rise (US Department of ings (such as ‘‘Smoking brings you and the people around you
Health, 2004). In an effort to help to reduce the number of tobacco severe damage” and ‘‘Smoking makes you unattractive”) are effec-
consumers, governments started several anti-smoking strategies, tive in reducing attitudes towards smoking to a greater degree the
such as raising the taxation on tobacco products, publicizing more people base their self-esteem on smoking. This is because
anti-smoking advertisements, or making certain public places such warnings may challenge the very reason for smoking particu-
smoke-free. larly for those who believe that smoking allows them to feel valued
Labeling of tobacco products is yet another anti-smoking strat- by others or to boost their positive self-image. Thus, a high smok-
egy. In many countries, every tobacco product has to have a mes- ing-based self-esteem may make people especially susceptible for
sage printed on the package that warns against the negative information that undermines their self-esteem.
consequences of smoking for one’s health. Such warnings shall Based on the terror management health model (Goldenberg &
make consumers aware that smoking leads to death (e.g., ‘‘Smok- Arndt, 2008), we predict on the other hand that the opposite pat-
ing kills”), to health problems (e.g., ‘‘Smoking clogs the arteries tern would emerge when the warnings are related to death and
and causes heart attacks and strokes”), and to social problems therefore make mortality salient. That is, the more individuals base
(e.g., ‘‘Protect children: Do not make them breath your smoke”). their self-esteem on smoking, the more they would adopt a posi-
Thereby, many of such warning messages remind us of our mortal- tive attitude towards smoking after being provided with mortal-
ity, but others do not (e.g., ‘‘Smoking makes your skin age ity-salient warning messages because awareness of mortality
quicker”). How effective are such warnings in reducing smoking motivates self-esteem striving (Pyszczynski, Greenberg, Solomon,
attitudes? Arndt, & Schimel, 2004). This hypothesis can be derived from terror
In the present research, we hypothesize that warning messages management theory (e.g., Greenberg, Solomon, & Pyszczynski,
may have different effects on smoking attitudes, depending on (1) 1997; Solomon, Greenberg, & Pyszczynski, 1991), which is based
how salient mortality is in the messages, and on (2) how strongly on the notion that all human beings are aware that their own death
recipients base their self-esteem on smoking. Specifically, we pro- is inevitable. This knowledge creates the potential for extreme anx-
iety (or terror) because of being helplessly exposed to this threat.
In order to manage this distress, people are motivated to maintain
* Corresponding author. Address: Department of Psychology, New York Univer-
sity, 6 Washington Place, New York, NY 10003, United States. faith in their cultural worldview and to keep a positive self-esteem
E-mail address: jochim.hansen@nyu.edu (J. Hansen). (i.e., they are motivated to increase the subjective belief that one is

0022-1031/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jesp.2009.09.007
J. Hansen et al. / Journal of Experimental Social Psychology 46 (2010) 226–228 227

a valuable member of one’s culture). Keeping a positive self-esteem ages my positive self-image” (reversed scored), ‘‘Smoking allows
can give a feeling of security and function to buffer people from the me to feel valued by others,” and ‘‘Smoking allows me to feel wor-
deeply rooted existential fear when mortality is made salient thy.” For each statement, participants indicated their agreement on
(Pyszczynski et al., 2004). To the extent that smoking is a source a scale ranging from 1 (do not agree at all) to 7 (agree completely).
of self-esteem, mortality-salient on-pack warnings would thus We calculated a mean smoking-based self-esteem index for each
ironically cause more positive attitudes towards smoking. participant (Cronbach’s a = .79). After this scale, participants
A comparable example for the buffer function of self-esteem has worked on a filler questionnaire.
been shown by an experiment that examined the effects of mortality Next, participants were randomly assigned to one of two condi-
salience on risk taking while driving (Taubman Ben-Ari, Florian, & tions. In both conditions we presented illustrations of two cigarette
Mikulincer, 1999; see also Jessop, Alberty, Rutter, & Garrod, 2008). packs, each of which contained an anti-smoking warning message.
Mortality salience inductions led to more risky driving than the con- In one condition, mortality was made salient by using warnings that
trol condition among individuals who perceived driving as relevant read, ‘‘Smokers die earlier” and ‘‘Smoking leads to deadly lung can-
to their self-esteem. Similarly, it has been demonstrated that wo- cer.” In the other version, the warnings were unrelated to death
men for whom their looks were an important source of self-esteem (i.e., ‘‘Smoking brings you and the people around you severe dam-
preferred to eat fruit salad (compared to chocolate cake) if mortality age” and ‘‘Smoking makes you unattractive”). A pretest with an inde-
was salient, whereas women who based their self-esteem on other pendent sample of 23 smokers had revealed that these two versions
things showed the opposite pattern (Ferraro, Shiv, & Bettman, 2005). were comparably threatening, t < 1, scary, t(21) = 1.843, p = .19, dis-
In the present research, we assessed smoking-based self-esteem turbing, t(21) = 1.902, p = .18, and self-relevant, t < 1, but differed in
and afterwards presented smokers with fear-evoking warning the degree to which they remind perceivers of the death as intended,
messages on cigarette packages. These warnings were either re- t(21) = 7.666, p = .01. In order to make sure that participants attend
lated to death or not. After a delay, we measured smoking atti- to the packages, they were asked to write down three thoughts that
tudes. We included a delay because any effects of mortality came to their mind when they saw the illustrations.
threats can only be found when death-related thoughts have been After this, participants filled in an unrelated questionnaire that
removed from conscious awareness, either by delay, by distraction, was included to introduce a 15-min delay between exposure of the
or by subliminal presentation of the mortality threats (Goldenberg warnings and the dependent measure. This was done because pre-
& Arndt, 2008; Greenberg, Pyszczynski, Solomon, Simon, & Breus, vious research has shown that effects of mortality salience emerge
1994). When thoughts of death are in focal attention, behavior is not until death-related thoughts have been removed from con-
guided by a motivation to overcome these thoughts. Only after scious awareness (Arndt, Greenberg, & Cook, 2002; Pyszczynski,
thoughts of death have become nonconscious, motivations to bol- Greenberg, & Solomon, 1999).
ster one’s self-esteem may override such proximal defenses. Subsequently, smoking attitudes were collected with five items
Because reminders of one’s mortality may lead to attempts to that were answered on respective 7-point scales (‘‘Do you enjoy
bolster one’s self-esteem in response to an existential threat, we smoking?” ‘‘How important is smoking for you?” ‘‘Do you intent
predicted that mortality salience would paradoxically cause more to smoke more or less in the future?” ‘‘Do you intent to quit smok-
positive smoking attitudes for individuals who build their self-es- ing in the future?” and ‘‘Are you going to smoke a cigarette directly
teem on smoking. Therefore, death-related anti-smoking warnings after this study?”). Finally, participants were thanked, debriefed,
should be less effective in changing smoking attitudes the more and given credit.
people base their self-esteem on smoking. In contrast, death–neu-
tral (but self-esteem related) anti-smoking warnings should more
Results
effectively induce anti-smoking attitudes the more smoking is a
source of self-esteem.
Preliminary analyses revealed that the mortality salience and
the control conditions did not differ significantly regarding
Method smoking-based self-esteem, (Mmortality salience = 2.25, SD = .84 vs.
Mcontrol condition = 2.32, SD = .81), t(37) = .28, p = .78, and smoking
Participants and design attitudes (Mmortality salience = 4.24, SD = 1.06 vs. Mcontrol condition =
3.93, SD = 1.00), t(37) = –.948, p = .349.
Thirty-nine psychology students who indicated that they were To test the hypothesis that smoking-based self-esteem moder-
smokers took part in the study in exchange for course credit (31 ates the effect of mortality salience on smoking attitudes and
female, 8 male). Age ranged from 17 to 41 years (M = 22.46, intentions, we regressed the attitude score on the experimental
SD = 4.603, Mdn = 21). They were randomly assigned to one of two condition (0 = mortality salience condition, 1 = control condition),
experimental conditions (mortality salience vs. control).1 The two the z-standardized smoking-based self-esteem score and their
conditions did not differ in the amount of smoking per day, two-way interaction. The main effect of condition was not signifi-
t(37) = 1.223, p = .26. cant, b = .16, t(36) = .94, p = .35, indicating a similar overall smok-
ing attitude in both conditions. Importantly, the interaction
between the experimental condition and the smoking-based self-
Materials and procedure
esteem was significant, b = .63, t(36) = 2.88, p < .01 (see Fig. 1). As
hypothesized, simple slope tests (Aiken & West, 1991) revealed
After providing some demographic data, participants filled in a
that the smoking-based self-esteem was negatively related to
series of questionnaires. The first questionnaire measured to what
smoking attitudes in the control condition, b = –.45, t(36) = –2.02,
degree participants based their self-esteem on smoking with 12
p = .05, but that it was positively related to smoking attitudes
items that were adapted from Taubman Ben-Ari et al. (1999),
when mortality was salient, b = .42, t(36) = 2.05, p < .05.
and included the following examples: ‘‘Smoking brings out un-
wanted aspects of my character” (reversed scored), ‘‘Smoking dam-
1
Discussion
Furthermore, we varied whether the warning messages contained pictures that
additionally emphasized the warning or not. However, this variation of the
presentation format had no effect on the attitudes, nor did it qualify the found The present findings suggest that warning messages on ciga-
effects. Therefore, it is not discussed further. rette packages can be effective in inducing anti-smoking attitudes.
228 J. Hansen et al. / Journal of Experimental Social Psychology 46 (2010) 226–228

to predict whether a death-related or a death-unrelated warning


message would be more effective. Yet, one could speculate that
certain populations base their self-esteem on smoking to a higher
degree than others, for instance young smokers who want to im-
press their peers. If this turns out to be true, a consequence of
our findings would be that such populations should be warned
against noxious consequences of smoking with death–neutral mes-
sages that undermine their smoking-based self-esteem. Such mes-
sages would probably not increase smoking attitudes as a strategy
to buffer against existential fears, but instead change people’s
minds after a delay.
In general, when smokers are faced with death-related anti-
Fig. 1. Attitudes towards smoking after anti-smoking warnings as a function of smoking messages on cigarette packs, they produce active coping
mortality salience of the warnings and smoking-based self-esteem (±1 SD). The attempts as reflected in their willingness to continue the risky
mean attitude score ranged from 1 to 7; higher values indicate a more positive smoking behavior. Which coping attempt they use, depends on
attitude towards smoking.
their smoking-based self-esteem. To succeed with anti-smoking
messages on cigarette packs one thus has to take into account that
However, their effect depends on a combination of smoking-based considering their death may make people smoke.
self-esteem and mortality salience of the message. On the one
hand, death-related warnings were not effective and even ironi-
Acknowledgment
cally caused more positive smoking attitudes among tobacco con-
sumers who based their self-esteem on smoking. This finding
We thank Malte Friese for helpful comments on an earlier ver-
suggests that individuals with a high smoking self-esteem use po-
sion of this article.
sitive smoking attitudes as a strategy to buffer against existential
fears provoked by the death-related warning messages. For indi-
viduals with a low smoking self-esteem, in contrast, a positive atti- References
tude towards smoking would not buffer against existential terror.
Thus, relative to high smoking self-esteem participants, partici- Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting
interactions. Newbury Park, CA: Sage.
pants with a low smoking self-esteem demonstrated lower smok- Arndt, J., Greenberg, J., & Cook, A. (2002). Mortality salience and the spreading
ing attitudes.2 activation of world-view relevant constructs: Exploring the cognitive
On the other hand, warning messages that were unrelated to architecture of terror management. Journal of Experimental Psychology:
General, 131, 307–324.
death effectively reduced smoking attitudes the more recipients Ferraro, R., Shiv, B., & Bettman, J. R. (2005). Let us eat and drink, for tomorrow we
based their self-esteem on smoking. This finding can be explained shall die: Effects of mortality salience and self-esteem on self-regulation in
by the fact that warnings such as ‘‘Smoking brings you and the peo- consumer choice. Journal of Consumer Research, 32, 65–75.
Goldenberg, J. L., & Arndt, J. (2008). The implications of death for health: A terror
ple around you severe damage” and ‘‘Smoking makes you unattrac-
management health model for behavioral health promotion. Psychological
tive” may be particularly threatening to people who believe the Review, 115, 1032–1053.
opposite, namely that smoking allows them to feel valued by oth- Greenberg, J., Pyszczynski, T., Solomon, S., Simon, L., & Breus, M. (1994).
ers or to boost their positive self-image. To the degree that warning Role of consciousness and accessibility of death-related thoughts in
mortality salience effects. Journal of Personality and Social Psychology, 67,
messages undermine the high smoking-based self-esteem, smok- 627–637.
ing may be devalued. Interestingly, this effect, too, only emerges Greenberg, J., Solomon, S., & Pyszczynski, T. (1997). Terror management theory of
when attitudes are assessed after a delay but not when attitudes self-esteem and cultural worldviews: Empirical assessments and conceptual
refinements. In M. Zanna (Ed.). Advances in experimental social psychology (Vol.
are assessed directly after the warning (see Footnote 2). Possibly, 29, pp. 61–139). San Diego, CA: Academic Press.
when smokers are consciously aware of warnings that argue Jessop, D. C., Alberty, I. P., Rutter, J., & Garrod, H. (2008). Understanding the impact
against the basis of their self-esteem, such warnings may be down- of mortality-related health-risk information: A terror management theory
perspective. Personality and Social Psychology Bulletin, 34, 951–964.
played. After a delay, however, the warnings are not in conscious Pyszczynski, T., Greenberg, J., & Solomon, S. (1999). A dual-process model of
awareness anymore and may unfold their impact. defense against conscious and unconscious death-related thoughts: An
In sum, consistent with the terror management theory (Green- extension of terror–management theory. Psychological Review, 106,
835–845.
berg et al., 1997; Solomon et al., 1991) and the terror management Pyszczynski, T., Greenberg, J., Solomon, S., Arndt, J., & Schimel, J. (2004). Why do
health model (Goldenberg & Arndt, 2008), the impact of warning people need self-esteem? A theoretical and empirical review. Psychological
messages on cigarette packs depended (1) on the degree to which Bulletin, 130, 435–468.
Solomon, S., Greenberg, J., & Pyszczynski, T. (1991). A terror management theory of
self-esteem was based on smoking, and (2) on the salience of death
social behaviour: The functions of self-esteem and cultural worldviews. In M.
in the warnings. Our finding is of high practical relevance, as it sug- Zanna (Ed.). Advances in experimental social psychology (Vol. 24, pp. 93–159).
gests that a differential strategy should be applied to warn smokers San Diego, CA: Academic Press.
against negative consequences of smoking, depending on the de- Taubman Ben-Ari, O., Florian, V., & Mikulincer, M. (1999). The impact of mortality
salience on reckless driving: A terror management mechanism. Journal of
gree to which they base their self-esteem on smoking. Death-re- Personality and Social Psychology, 76, 35–45.
lated warnings are not effective and even have unwanted effects US Department of Health and Human Services (2004). The health consequences of
when smokers have a high smoking-based self-esteem. smoking: A report of the surgeon general. Atlanta, GA: US Department of Health
and Human Services, Centers for Disease Control and Prevention, National
However, outside the laboratory, the degree to which smokers Center for Chronic Disease Prevention and Health Promotion, Office on Smoking
base their self-esteem on smoking is unknown. Thus, it is difficult and Health.

2
Please note that this effect only emerged when attitudes had been assessed after a
delay. In a different study in which attitudes were measured directly after the
manipulation of mortality salience, we did not find any evidence for the influence of
mortality salience, smoking-based self-esteem, and their interaction on smoking
attitudes, ts < 1. This finding is in accordance with the terror management health
model (Goldenberg & Arndt, 2008).

You might also like